Dogma wrote:until that testing is done, we are saddled with existing dogma.
If we focus on the concept, I feel we this is more an issue of common sense rather than "dogma". And if we divert to semantics, we are only playing word games.
We can strive to account for that "dissimilarity/discrepancy/difference" when we are confronted with one, but I think everyone would agree that it is impossible for two different areas of friction skin to have a large number of level two points (say, 30) that "match." Consequently, when that large number of points are found to match (please, let's don't play semantics with my use of that term -- you know what I mean) we believe the two prints to have been made by the same person, in spite of that one dissimilarity. The problem lies in deciding what is that magic number beyond which no two areas of skin will have matching points.
Has anyone else out there tried to trick AFIS into finding an erroneous identification? I have. I have played with fingerprinting folks in the lab and entering 5 or 6 very "common" looking points for an AFIS search. For example, I might pick a half dozen ridge endings pointing upwards along either side of the type lines above a delta, and enter only those 6 points. Of course, I am using a fully rolled inked print, but I'm only entering those six points to search. I have been amazed at some of the close matches, going well beyond the five or six points I have searched. Try running twenty or thirty prints and entering only a few points in common groupings, not unusual clusters but very common looking clusters.
The point is this -- I have found a few close matches with as many as 12 or 14 points that could conceivably match, except for maybe one dissimilarity in the middle. Now, using inked prints there is no danger of making that erroneous identification. But what if a latent print included only those 12 or 14 matching points and that one dissimilarity, maybe from a distorting curved surface, say a beer bottle or a drinking glass with a little condensation that masked out all the level three detail? (Anybody seen the prints in the Cowans case out of Boston? What about the Mayfield case?)
Back in the early 1980's I read a short editorial letter, either in "Identification News" or in the Texas Division newsletter, in which the author took the position that we need to apply a higher standard of identification in AFIS searches than in comparisons that result from the detective naming a suspect for us to compare. I remember thinking what nonsense that was because all identifications have to meet the same very high standard of absolute certainty -- identification to the exclusion of any other person who is alive, ever has lived, or ever will live. So what was this writer thinking? Well, having seen cases like Cowans and Mayfield, I am beginning to understand. And as I travel and teach, almost everywhere I go somebody has another example from their AFIS of another "close non-ident" with ten or more points that "match."
So, to circle back to the "one discrepancy rule," I still ponder what is that high number of points beyond which we can be certain of an identification when there is that one dissimilar point somewhere in the latent?
So, to circle back to the "one discrepancy rule," I still ponder what is that high number of points beyond which we can be certain of an identification when there is that one dissimilar point somewhere in the latent?
Dogma wrote:
The point I am trying to make is that without being able to assign clear, quantifiable and tangible values to concepts like sufficiency and thresholds......
and
We can no doubt test for the theory that an overwhelming (or specified) amount of detail in agreement between two prints nullifies a dissimilarity (as defined by you and not SWG).
It seems that there is that threshold that allows an examiner to call a "match" absent any exclusionary type differences. However, when such differences are present, the threshold for a "match" would be much higher, giving us a situation akin to the quality/quantity pyramid. Now if we could only decide what the value of that magic threshold is.
In my first posting I predicted this thread would revolve around certain criteria. We appear to have come to the phase of separating quantity from quality. My next prediction is that quantifying alone will not substantiate the 'one discrepancy rule". I think pat would agree that a valid identification can be made using all three levels of detail even when the second level is minimal in quantity. Perhaps its like a game of cricket where there are three stumps and two bails. You're 'out' if either of the two bails are removed. The stumps could represent the three levels of detail and I guess the bails could be their quality and quantity. If the whole set is safe then you're still in the game. Who'se got the ball?
Ah, yes, Les -- this brings to mind the old mantra, "Show me the print!" In the latent prints we look at, it is virtually impossible for our minds to separate level 2 quantity from level 3 quality before we reach our conclusions. Likewise, it is very hard to discuss hypotheticals when there are literally an infinite number of variables.
On a note somewhat related to the topic of this thread, the second fingerprint lecture at the IAI in Boston was a talk by Cedric Neumann and Paul Chamberlain (absent was Christophe Champod). Cedric and Paul presented research they are doing with Christophe into numbers of points at level two and an associated measure of reliability for an identification with the specific points in the latent. In other words, probability. One of the issues that my poor memory seems to recall was an effort to assess a probability when there is a dissimilarity in conjunction with an otherwise acceptable number of points. Of course, their model ignores level 3 detail.
As you might have anticipated, there was some rather hostile questioning punctuated with negative comments following their presentation. But I was quite encouraged by their presentation. Let me say that I don't think it would be the end of our science if a probability model were developed. It would be one more tool for us to use in our science toolbox. They have taken the concept of reliability or probability to levels I never thought possible and I look forward to seeing their refinements. I remember a conversation with Christophe years ago in which he forecast the day when a key on the computer would allow us to calculate a probability in conjunction with an AFIS search. I laughed and told him I did not think that day would ever come, but I wished him luck. Hearing from Cedric and Paul what they have accomplished makes me think Christophe was probably right (forgive the pun) and we will someday have a more or less reliable methodology of assessing reliability of a specific identification.
No where would this tool be more valuable to us than in the situation where we have plenty of good detail, but one pesky "unexplainable dissimilarity."
Forgive me if I've misunderstood the Inspector General report on Mayfield, but didn't it recommend some basic studies on whether Level 3 is truely permanent and to what extent it should be relied on in making matches? In the absence of that work, how solid is the Level 3 stump (to use Les' analogy)?
I kind of short changed you by offering what you said as tongue in cheek as it is more than that as usual. You are right, you do make me think and I can read your post 2-3 times and understand it without having to put in under my pillow at night. I will not short change you again.
You used the word dogmatic, and I have seen it several times on this site. I thought I knew what it meant but I looked it up anyway. Wow—Bad Word, “Characterized by an authoritative, arrogant assertion of unproved or un-provable principles”. I think when people use that word it is more of an accent to a particular point or belief rather than the adjective that it is. I believe there are a number of people out there with different viewpoints on how things should or should not be done which were developed over a period of time within their sphere of experiences and beliefs. I may not agree with those beliefs any more than they believe in mine. I will have to admit though that some of them are very assertive at times.
I agree with you that the comparison aspect should not be blind but should take in other aspects such as distortion, disruptions, matrix, substrate, and methods of development, if it is viable to the analysis. In my world we deal mainly with property crimes. If I have a homicide with an axe and I have a marginal print on the handle, you bet all those factors better play into the analysis of the latent. But if it is from the exterior window glass of vehicle in a burglary of a vehicle and the print is very good with 35-40 clear points and no ananomly, that information about distortion, matrix, substrate and MOD do not mean a twit. It is about context.
On the doctor’s analogy, you are right in that I would not want my doctor to base their analysis on a single symptom anymore than I would a latent print examiner on a single event. I would hope they would not. My analogy is a little different.
I have a big tree in my back yard and I want to know what kind of tree it is. Having considered a degree in botany in college I think I will take a stab at it. I grab my book by Robert Vines “Trees of Central Texas” and go out back. I think this might be a Post Oak. I look at the bark and it is reddish brown, thick, divided into regular fissures with plate like scales, just like the standard in the book. I pull a limb down and look at the leaves. Yes they are simple, alternate, deciduous, and oblong-obviate blades 4-7 inches long with deep rounded sinuses with the base cuneate or rounded. But wait a minute I see another leaf and it is different from the standard. It is oblong –lanceolate with the blade 2 inches long and serrate on the margin. What caused this leaf to appear like it does? Was it genetic? Did disease cause the anomaly or insect/animal destruction? Or was it a small piece of ice falling from a spring sky when the leaf was first forming. I do not know. Will it make a difference in my determination if this is a Post Oak or not. No it will not. Anomalies are a part of natural life and the world we live in. If a vast amount of information points us in one direction, why should a single anomaly deter us from that course? I know this is dangerous ground to a new examiner, I know that it is a departure from what was before, I know it has problems, but in all honesty as an analyst I cannot throw out the baby with the bathwater. To ignore the overwhelming information that is presented for a single thing out of place is not a good reasonable procedure. Or the only way I can count the information is to know exactly what caused it OR make up an excuse for what caused it. On some of the anomalies that I have seen in the past 37 years I do not think anyone can say exactly what caused it. I can give you maybe two or three reasonable answers to the delimina but not the exact one. I would rather give two or three reasonable answers than to say well it was the powder, or the pressure, or the movement, or they did not lift it right, or it’s a double tap or whatever just so I can claim the identification. I think the issue in the beginning was why do we have to know exactly what caused it as scientist (yes I used that word) to be able to use the information.
One more topic and then I will leave you alone. On your post after mine in the fourth paragraph you say “Also, certain “points” are rarer than others”. Now that is a hot topic with me and you want to talk about passion. Rarity, Locard talked about it. It has been a part of my consideration in a number of very tough latent prints in the past 37 years. In my first fingerprint school with the FBI there was an outline with the concept of “rarity” on the front page. Went back a few years later and given a different handout and the concept of rarity is on the back page. Went back a decade later and the word had been exorcised from the vocabulary. When I was in SWGFAST, they were dealing with (when I say they I mean another committee within SWGFAST) Comparison Methodology. They gave a brief presentation as to how far they had gotten. During a break I asked some if they were going to broach the subject of rarity. They looked at me and said “Rarity---Rarity, we don’t even want to go there”. I was shocked, and for two days I was speechless (I know that is hard to believe, but I was). Here assembled are the greatest minds in fingerprints and they do not want to broach the subject of rarity. I got home and thinking that I had this all wrong all the time. It is all about unique and permanent and rarity does not have a thing to do with it. I thought and went back to reading and thinking that my concepts needed to change. But the more I thought and the more I looked and the more I read, I realized that I was not wrong, they were wrong. Rarity can be a very important aspect of an analysis or comparison. But now I see a word that is being used more and more. That word is weight. That word weight is starting to appear all over the place on this website and in writing. Well this has more weight than that. This should be judged more as the uniqueness of it has more weight. Well I got news for you folks that is rarity. If it walks like a cat, got fur like a cat, and meows like a cat why don’t you just call it a cat. But they do not want to do that because the word rarity does not fit too well in the glove called unique. How can something be rare and unique? I guess not the same way as something having weight and also unique. (Is not weight generally associated with a number?). By the way you will not see the concept of weight in SWGFAST documents. I guess it has not got there yet. But the documents put out by by the IEFFG discuss it. Shoot they even have pictures. The writings of Locard are out the window.
Forgive my passion on this subject Dogma as I know it is not part of your experience. I have more to say but I need to go home. Home wins out. Keep up the good fight and your reasoning.
By the way I have quit using the word dissimilarity to mean an explainable difference. I am now using the word ananomly. One must conform.
Lisa -- while the use of "level 3" details is being called into question as you point out, there is still weight one can reasonably give to such details during a comparison. For example, if I see a prominent bump on a ridge in a latent print, say two ridge units down from a ridge ending, and there in the inked print is a similar bump on the ridge in the same position with regard to that ridge ending, my mind cannot help but count that of some value or weight toward the identification. On the other hand, if there is a prominent, unmistakeable bump on the ridge in the latent and no anomoly at all in the inked print, that fact might introduce doubt into the comparison.
Charles -- I'm using some of your words, too. In my class, I taught that with a set standard, say "16 points," it was like putting little one ounce brass weights on a scale. Each little brass weight is exactly the same and when you have 16 ounces (points), you have a pound (identification) and the scale tips. In reality, the human mind during fingerprint comparison takes features into consideration without exact weights. Some features are like dropping a pebble onto the scales, some have no more weight than a grain of sand, and some are like dropping a cobblestone on the scale -- thump! What about that anomoly (dissimilarity)? I would say that is a pebble of some weight dropped onto the other weighing pan of the scale and it has to be counterbalanced with extra features on the pan labelled "identification." But still, when the scale tips, your mind has the "Ah ha!" moment and you claim the identification, or at least your mind has developed that "bias" and you are close to claiming it.
Okay, my metaphors and analogies are carrying me away. Time for my second cup of morning coffee.
I saw an infomercial for the Dogmatic on late night TV last week. It's a really neat looking machine that looks like a doghouse. You send your pet through it for a bath. But seriously, folks, . . . . .
One senior latent print examiner I know uses the word "dogma" when referring to the philosophy of point counting and the use of a threshold for identification. Another senior latent print examiner I know uses the word "dogma" when referring to concepts espoused by David Ashbaugh and other proponents of ridgeology. Both of these people are well known, widely respected experts whom I consider good friends of mine. But the word "dogma" is a dart thrown back and forth between the two, as if it would wound each other's arguments. (Okay, so the second cup of coffee renewed the impulse to use metaphors.)
It seems to me there is validity in reliance on some dogmas. Other dogmas fly in the face of reason and need to be examined and tested. But the word has been ill used by many on both sides of not only fingerprint issues, but just about all controversies. Consequently, it has become a turnoff for me and almost has an effect opposite of that intended by a person who uses the word to discredit another's arguments.
So, Dogma, it seems to me your name is used rather loosely by people on both sides of an argument to describe the opposing beliefs in an attempt to bias listeners (or readers) against them.
I wonder what the bonus gift is if I call the toll-free number and order my Dogmatic in the next thirty minutes? But WAIT !!! Be among the first five hundred callers and . . . . . a free bottle of Catatonic?
You have probably seen this but there is a new show in the fall called SHARK. About a defense attorney who goes to work for the DA. It stars James Woods.
In one clip he is in an office with several people and someone says:
I had a good discussion one time with a defense attorney who understood fingerprints very, very well. He explained one of his strategies (I forget what it was) that left the jury with a very wrong impression of fingerprint evidence. I commented, "But that's not the truth!" His reply has always stuck with me. He said, "The court is not concerned with 'truth.' It is only concerned with 'fact,' and a 'fact' is anything I can prove to a jury, whether it's the truth or not."
Pat, without more detail, that raises some hairs on the back of my neck. An attorney has an obligation of candor to the court under the ethics rules. That doesn't mean the defense has to help the state by showing all of its cards. Doesn't mean the defense can't test the state's burden of proof with a withering cross-exam, even if we strongly feel the evidence is solid. But if an advocate is arguing in a motion or closing argument something he or she knows to be false -- that worries me.
I think there’s plenty of cases where attorney’s purposely give false information (or they twist the truth to suit their needs) and they do it so subtly that if their caught they can claim it was simply a misunderstanding.
Take for example in US v Byron Mitchell. The “memorandum of law in support of Mr. Mitchell’s motion to exclude the government’s fingerprint evidence” refers to the erroneous identification in the Andrew Chiory case by referring to an article written by Stephen Gray. The motion states, “In that case, two latent prints that had been recovered from a burglary crime scene were each found to have at least sixteen points in common with two of the defendant’s inked prints. These identifications, pursuant to standard Scotland Yard procedures, had been triple checked prior to the defendant’s arrest.”
There’s two false statements here. First, the article by Stephen Gray never states that 2 latent prints each had 16 points in common. The article states, “the standard requires at least 16 points”. The article never states this standard was met. The second part that’s twisted is when it says “These identifications, pursuant to standard Scotland Yard procedures, had been triple checked prior to the defendant’s arrest”. The article doesn’t state this either. It states that the prints were allegedly triple-checked.
These are small changes that might seem insignificant but the impact they could have on a judge, a jury, or on other court cases could be huge. (I mention other court cases because I’m assuming that it’s common for attorney’s to take this information from this motion and use it in motions for other cases, without verifying whether or not it’s true).
By stating that attorney regularly act this way, I’m not trying to imply that attorney’s are the only ones who do this. I could probably give you just as many examples of people from the fingerprint industry that’ve done the same thing. Some of it may be intentional and some of it may be lack of information (that’s sort of what this thread is all about), but either way I’m with you…..it raises some hairs on the back of my neck.
Michele
The best way to escape from a problem is to solve it. Alan Saporta
There is nothing so useless as doing efficiently that which should not be done at all. Peter Drucker
(Applies to a full A prior to C and blind verification)
Whoever is looking into the validity of third level detail should be visiting the bibliographies of several authors who use this site including my own. Third level detail is as solid as the other two. In 'fact' all three are interdependent and biologically bound through embryology, physiology and anatomy. Scars would fall into a different subcategory of third level because of their set of variables.
I too have had a defense lawyer say outside of court, ' we are not interested in the truth'. Its a bit like our system of parliamentary privilege but is consequence can be very harmful for victims of crime.
The issue of dogma is one that should be resolved whenever the SWGFAST team complete all their portfolios. Dogma will eventually be replaced by guidelines and rules. Once we all agree to the interpretation of what has been transcribed/converted then we move forward on a stronger basis for our science. Even today the Justice 1 Committee into McKie is hearing expressions that fingerprints are a 'police' science. While I can agree the majority of fingerprint collections are administered under a police management system the science of fingerprints stands within the general scientific community. Cheers.
